Customer relationship management system and method based on artificial intelligence

Abstract

The invention relates to the technical field of customer management, in particular to a customer relationship management system and method based on artificial intelligence. A customer portrait corresponding to each target customer is constructed based on customer information, then a management model is created, a customer standard portrait is screened out based on the customer portraits through the management model, user levels of target users are set based on the customer standard portrait, and the trained management model is obtained. And finally, inputting the real-time information into the trained management model, obtaining a real-time standard portrait of the real-time customer, and predicting and adjusting the customer relationship of the real-time customer in combination with the real-time standard portrait. Therefore, the reliability of customer relationship management is improved, and the customer relationship of the target customer in the future time threshold is predicted through the customer standard portrait, so that the user can prevent the target customer in advance, and risk event exposure of the target customer is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of customer information, in particular to a customer relationship management system and method based on artificial intelligence. BACKGROUND

[0002] Customer information refers to customer preferences, customer segmentation, customer needs, customer contact information and other basic information about customers. Customer information is mainly divided into three types: descriptive information, behavioral information and associated information. Scientific customer information management is an important guarantee for customer aggregation and business development. Due to the characteristics of customer information, scientific customer information management is required for information processing, information mining, information extraction and reuse. Through customer information management, the maximization and optimization of customer information utilization can be achieved.

[0003] A customer information processing method and device based on customer portrait are disclosed in Chinese Patent No. CN114066620A. The method includes taking pre-stored customer information, which includes multiple customer portrait indicators. The customer portrait indicators are vector standardized according to the indicator values of the customer portrait indicators to obtain the correlation coefficient matrix, eigenvalues and eigenvectors of the customer portrait indicators. The contribution rate of each indicator value is determined using the eigenvectors and contribution rate algorithm, and the principal component indicators are selected from the indicator values according to the preset contribution rate threshold. The comprehensive indicator score of the customer portrait indicators is determined according to the eigenvalues, eigenvectors and principal component indicators, and the comprehensive indicator score is evaluated and graded using a preset grading rule to obtain multiple evaluation grades for business processing. However, the existing technology only stops at constructing customer portraits and does not further predict and analyze customer relationships based on customer portraits, making it difficult for users to predict customer relationships in a timely manner and leading to difficulties in adjusting customer relationships with target customers in a timely manner. SUMMARY

[0004] The present application aims to solve the problems in the background art and provides a customer relationship management system and method based on artificial intelligence.

[0005] The technical solution of the present application is as follows: On the one hand, the present application provides a customer relationship management method based on artificial intelligence, which includes: Collecting customer information of multiple target customers and constructing a customer portrait corresponding to each target customer based on the customer information; Creating a management model; Inputting the customer information and the customer portrait corresponding to each customer information into the management model, filtering out a customer standard portrait based on the customer portrait through the management model, setting a user level of the target user based on the customer standard portrait, and obtaining a trained management model; Collect real-time information of real-time customers, input the real-time information into the trained management model, obtain the real-time standard profile of the real-time customers, and predict and adjust the customer relationship of the real-time customers based on the real-time standard profile.

[0006] Preferably, customer information of multiple target customers is collected, and a customer profile corresponding to each target customer is constructed based on the customer information, including: Create a customer database; Collecting customer information of multiple target customers respectively, and inputting all collected customer information into a customer database; the customer information includes basic customer data and customer behavior data, the basic customer data includes customer address and customer contact information, etc., and the customer behavior data includes transaction record data and customer communication data; Combine the customer information of each target customer to build a corresponding customer portrait.

[0007] Preferably, a corresponding customer profile is constructed based on the customer information of each target customer, including: Randomly select customer information of a target customer from the customer database; Calculate the customer value of the target customer using Formula 1 based on the customer information of the target customer; Formula 1; in, is the customer value of the target customer, is the target customer’s i-th customer sub-information, is the weight corresponding to the i-th customer sub-information, and N is the total number of customer sub-information contained in the customer information; Return the customer information of a randomly selected target customer from the customer database until the customer information of all target customers in the customer database is selected, and obtain the customer value of each target customer; Set customer profiles and the customer value range corresponding to each customer profile; Assign each customer to the corresponding customer profile based on the customer value of the target customer.

[0008] Preferably, the customer information and the customer profile corresponding to each customer information are input into the management model, the management model is used to filter out the customer standard profile based on the customer profile, and the user level of the target user is set based on the customer standard profile, to obtain a trained management model, including: Constructing a coupling relationship between customer information and customer profile, and using customer information, customer profile, and the coupling relationship between customer information and customer profile as a training sample to obtain multiple training samples; Divide multiple training samples into training sets and test sets according to random proportions; inputting the training samples in the training set into the management model in sequence, performing clustering analysis on the multiple customer information under each customer portrait based on the customer portrait through the management model, so as to obtain a customer standard portrait corresponding to each customer portrait, and then setting the customer relationship of the target customer based on the customer standard portrait; inputting the test set into the trained management model to verify whether the trained management model is trained.

[0009] Preferably, the training samples in the training set are inputted into the management model in sequence, clustering analysis is performed on the multiple customer information under each customer portrait based on the customer portrait through the management model, so as to obtain a customer standard portrait corresponding to each customer portrait, and then the customer relationship of the target customer is set based on the customer standard portrait, which comprises: For each customer portrait, a training sample under the customer portrait is randomly selected from the training set, and the selected training sample is recorded as an initial standard portrait; other training samples are assigned to the initial customer portrait of the same customer portrait, thereby forming multiple portrait clusters; a threshold of iteration times is set; For each cluster, the distance between each training sample in the cluster and the standard portrait is calculated respectively, and the point corresponding to the average value of the distance is recorded as a new standard portrait; whether the iteration times is greater than or equal to the threshold of iteration times is judged; when the iteration times is greater than or equal to the threshold of iteration times, the iteration is stopped and the standard portrait obtained in the last time is recorded as the customer standard portrait, thereby obtaining the customer standard portrait corresponding to each customer.

[0010] Preferably, the training samples in the training set are inputted into the management model in sequence, clustering analysis is performed on the multiple customer information under each customer portrait based on the customer portrait through the management model, so as to obtain a customer standard portrait corresponding to each customer portrait, and then the customer relationship of the target customer is set based on the customer standard portrait, which further comprises: a plurality of customer grades and a customer standard portrait corresponding to each customer grade are set; specifically, each customer grade can correspond to multiple customer standard portraits; a corresponding customer grade is given based on the customer standard portrait corresponding to each target customer.

[0011] Preferably, real-time information of real-time customers is collected, the real-time information is inputted into the trained management model, a real-time standard portrait of the real-time customer is obtained, and the customer relationship of the real-time customer is predicted and adjusted in combination with the real-time standard portrait, which comprises: real-time information of real-time customers is collected; a real-time portrait of the real-time customer is created based on the real-time information; input the real-time information and the real-time portrait into the trained management model to obtain a real-time customer corresponding customer standard portrait output by the trained management model; and record the real-time customer corresponding customer standard portrait as a real-time standard portrait; predict and adjust the customer relationship of the real-time customer in combination with the real-time standard portrait.

[0012] Preferably, the method for predicting and adjusting the customer relationship of the real-time customer in combination with the real-time standard portrait comprises: setting a time threshold; obtaining a future standard portrait corresponding to the time threshold of the real-time standard portrait; judging whether the future standard portrait and the real-time portrait are the same customer portrait; if the real-time standard portrait and the real-time portrait are not the same customer portrait, assigning the future standard portrait to the real-time customer; setting a corresponding customer relationship according to the future standard portrait assigned to the real-time customer.

[0013] In another aspect, the application further provides a customer relationship management system based on artificial intelligence, comprising a collection component and a management component, wherein the customer information of a plurality of customers is collected through the collection component, the customer relationship management method based on artificial intelligence is executed through the management component, the customer portrait is constructed based on the customer information through the management component, and the customer portrait is compared with the customer standard portrait, so as to adjust the customer relationship of the customer.

[0014] Preferably, the collection component comprises a collection module and a security module, the collection module is used to collect the customer information of a target customer, and the security module encrypts the customer information to prevent the customer information from being leaked.

[0015] Compared with the prior art, the above technical solution of the application has the following beneficial technical effects: By collecting the customer information of a plurality of target customers, constructing the customer portrait corresponding to each target customer based on the customer information, then creating a management model, and inputting the customer information and the customer portrait corresponding to each customer information into the management model, the customer standard portrait is screened out based on the customer portrait through the management model, the user level of the target user is set based on the customer standard portrait, the trained management model is obtained, finally the real-time information of a real-time customer is collected, the real-time information is input into the trained management model, the real-time standard portrait of the real-time customer is obtained, and the customer relationship of the real-time customer is predicted and adjusted in combination with the real-time standard portrait, the customer standard portrait is established in the application, and the customer standard portrait is used as the evaluation standard of the customer relationship, so as to improve the reliability of the customer relationship management, and the customer relationship of the target customer within a future time threshold is predicted through the customer standard portrait, so that the user can prevent the target customer in advance to avoid the exposure of the target customer to a risk event. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a customer relationship management method based on artificial intelligence proposed by the present invention; Figure 2 This is a principle block diagram of an artificial intelligence-based customer relationship management system proposed by the present invention; Description of the drawings: 100, acquisition component; 101, acquisition module; 102, encryption module; 200. Management components. DETAILED DESCRIPTION

[0017] Example 1, as Figure 1 As shown, the present invention proposes an artificial intelligence-based customer relationship management method, comprising: S100, collecting customer information of multiple target customers, and building a customer profile corresponding to each target customer based on the customer information; S200, creating a management model; S300: Inputting customer information and a corresponding customer profile of each customer information into a management model, filtering a standard customer profile based on the customer profile using the management model, and setting a user level for the target user based on the standard customer profile, thereby obtaining a trained management model; S400 collects real-time information of real-time customers, inputs the real-time information into the trained management model, obtains the real-time standard portrait of the real-time customer, and predicts and adjusts the customer relationship of the real-time customer based on the real-time standard portrait.

[0018] In the present invention, by collecting customer information of multiple target customers, a customer portrait corresponding to each target customer is constructed based on the customer information, and then a management model is created, and the customer information and the customer portrait corresponding to each customer information are input into the management model. The management model is used to screen out customer standard portraits based on the customer portraits, and the user level of the target user is set based on the customer standard portraits to obtain a trained management model. Finally, real-time information of real-time customers is collected, and the real-time information is input into the trained management model to obtain real-time standard portraits of real-time customers. The customer relationship of the real-time customer is predicted and adjusted in combination with the real-time standard portraits. This application improves the reliability of customer relationship management by establishing customer standard portraits and using customer standard portraits as evaluation criteria for customer relationships, and predicts the customer relationship of target customers within a future time threshold through customer standard portraits, so that users can take precautions against target customers in advance to avoid exposure to risk events of target customers.

[0019] In an optional embodiment, the step S100 includes: S110, creating a customer database; S120, respectively collect customer information of a plurality of target customers, and input all collected customer information into a customer database; the customer information includes customer basic data and customer behavior data, the customer basic data includes customer address and customer contact information, etc., and the customer behavior data includes transaction record data and customer communication data; S130, constructing a corresponding customer portrait combined with the customer information of each target customer; Specifically, before constructing the customer portrait, the keywords in the customer information need to be extracted, and the extraction of the keywords can be completed through the BERT model. After extracting the keywords, the keywords are standardized and converted into keyword vectors, thereby facilitating the subsequent creation of the customer portrait. Optionally, before constructing the customer portrait, the customer information can be preprocessed, including deduplication and leakage checking, so as to remove invalid data in the customer information and improve the training efficiency of the management model.

[0020] It should be noted that the core innovation of the BERT model is to capture the context information of the text through the bidirectional Transformer encoder, which breaks the limitation of the traditional NLP model of one-way processing.

[0021] In the present application, the customer information of a plurality of different target customers is collected, and a corresponding customer portrait is created for each customer based on the customer information. The customer portrait is a tool for sketching the target customer, and the characteristics of the target customer are reflected through the customer portrait, that is, the customer portrait is representative, so that the user can directly obtain the importance level of the target customer when managing the customer relationship through the customer portrait, thereby improving the management efficiency.

[0022] In an optional embodiment, the S130 comprises: S131, randomly selecting customer information of a target customer from the customer database; S132, calculating the customer value of the target customer based on the customer information of the customer through formula 1; Formula 1; Wherein, is the customer value of the target customer, is the i-th customer sub-information of the target customer, is the weight corresponding to the i-th customer sub-information, and N is the total number of customer sub-information contained in the customer information; Specifically, the more types of sub-data information contained in formula 1 when calculating the customer value, the more accurate the construction of the customer portrait is. S133, return the customer information of a target customer randomly selected from the customer database until the customer information of all target customers in the customer database is selected, and obtain the customer value of each target customer; S134, set the customer portrait and the customer value range corresponding to each customer portrait; S135, distribute each customer to the corresponding customer portrait according to the customer value of the target customer.

[0023] It should be noted that the customer value is actually an index for describing the importance of the target customer. Generally speaking, the greater the customer value of the target customer, the greater the corresponding customer value, and therefore, the higher the customer grade, the greater the range of the corresponding customer value. For customer grades, one customer grade can correspond to multiple different customer portraits, such as customer portrait A "export enterprise with annual output value of 1 million yuan" and customer portrait B "inland enterprise with annual output value of 200 million yuan". The importance of the user is equal, so customer portrait A and customer portrait B can be set as the same customer relationship.

[0024] In an optional embodiment, the S300 comprises: S310, construct the coupling relationship of customer information-customer portrait, and take the customer information, customer portrait, and coupling relationship of customer information-customer portrait as a training sample to obtain multiple training samples; S320, divide the multiple training samples into a training set and a test set according to a random ratio; S330, input the training samples in the training set into the management model in sequence, perform clustering analysis on the multiple customer information under the customer portrait based on the management model, to obtain the customer standard portrait corresponding to each customer portrait, and then set the customer relationship of the target customer based on the customer standard portrait; S340, input the test set into the trained management model to verify whether the trained management model is trained. Specifically, when verifying whether the trained management model is trained, the response time or output accuracy of the trained management model can be used as a judgment standard.

[0025] It should be noted that the present application takes the customer information and the customer portrait as the training data of the management model, thereby clustering the customer portrait through the management model, dividing the customer portraits of different target customers into different types of customer portrait clusters according to the customer information, taking the cluster center of each customer portrait cluster as the customer standard portrait of the target customer corresponding to this type of customer portrait, thereby establishing a standard evaluation system for the target customer, and finally setting the customer relationship based on this evaluation system to ensure the effectiveness and reliability of the customer relationship.

[0026] In an optional embodiment, the S330 comprises: S331, for each customer portrait, randomly selecting a training sample under the customer portrait from the training set, and recording the selected training sample as an initial standard portrait; S332, assigning other training samples to the initial customer portrait of the same customer portrait, thereby forming a plurality of portrait clusters; S333, setting a threshold of iteration times; S334, for each cluster, respectively calculating the distance between each training sample in the cluster and the standard portrait, and recording the point corresponding to the average value of the distance as a new standard portrait; S335, judging whether the iteration times are greater than or equal to the threshold of iteration times; S336, when the iteration times are greater than or equal to the threshold of iteration times, stopping iteration and recording the standard portrait obtained in the last time as a customer standard portrait, thereby obtaining the customer standard portrait corresponding to each customer; Specifically, the present application clusters all customer portraits through the K-means clustering method, which is a widely used clustering algorithm and aims to divide a data set into K clusters so that each data point belongs to the nearest center point.

[0027] It should be noted that the present application clusters all customer portraits, thereby screening out customer standard portraits corresponding to different customers, so as to unify the customer portraits corresponding to customers of the same type, i.e., establishing portrait standards of different types of customers through customer standard portraits, so that subsequent customer relationship management of real-time customers can be performed under the same standard, thereby ensuring the reliability of customer relationship management.

[0028] In an optional embodiment, the S300 comprises: S337, setting a plurality of customer grades and customer standard portraits corresponding to each customer grade; specifically, each customer grade can correspond to a plurality of customer standard portraits; S338, based on the customer standard portrait corresponding to each target customer, assigning a corresponding customer grade.

[0029] It should be noted that after establishing portrait standards corresponding to different types of customers through customer standard portraits, different customers can be divided into customer grades based on the customer standard portrait, thereby improving the efficiency of customer relationship management.

[0030] The existing customer standard portrait A and customer standard portrait B, the customer type corresponding to the customer standard portrait A is a customer type engaged in industrial production related, and the customer type corresponding to the customer standard portrait B is a customer type engaged in service industry related. The customer levels set for the customer standard portrait A and the customer standard portrait B are 1 level and 2 level respectively. Therefore, when the customer level of the customer engaged in the industrial production related is divided in the subsequent stage, the target customer can be quickly divided into the 1 level customer.

[0031] In an optional embodiment, the S400 comprises: S410, collecting real-time information of a real-time customer; S420, creating a real-time portrait of the real-time customer based on the real-time information; S430, inputting the real-time information and the real-time portrait into the trained management model to obtain a customer standard portrait corresponding to the real-time customer output by the trained management model; and recording the customer standard portrait corresponding to the real-time customer as a real-time standard portrait; S440, predicting and adjusting the customer relationship of the real-time customer in combination with the real-time standard portrait.

[0032] It should be noted that, since the trained management model is obtained in the foregoing step, after the real-time information of the real-time customer is collected, the trained management model can be used to quickly perform cluster analysis on the real-time portrait of the real-time customer, so as to find a customer portrait cluster in which the real-time portrait of the real-time customer is located. The customer standard portrait corresponding to the cluster center of the customer cluster is the real-time standard portrait of the real-time customer, so that the customer level division of the real-time customer is quickly completed.

[0033] In an optional embodiment, the S440 comprises: S441, setting a time threshold; S442, obtaining a future standard portrait corresponding to the time threshold of the real-time standard portrait; S443, judging whether the future standard portrait and the real-time portrait are the same customer portrait; S444, if the real-time standard portrait and the real-time portrait are not the same customer portrait, assigning the future standard portrait to the real-time customer; S445, setting a corresponding customer relationship according to the future standard portrait assigned to the real-time customer.

[0034] It should be noted that, since the center cluster obtained by the cluster analysis has the ability to represent all target points in the cluster, the real-time customer portrait has the ability to represent the real-time customer level. Therefore, the customer relationship of the real-time customer can be adjusted by using the real-time customer portrait.

[0035] The application predicts the change of the customer level of the real-time customer by using the change of the customer level of the real-time customer portrait at a future time threshold, thereby preventing the occurrence of customer risk event exposure and the like, and protecting the property safety of the user.

[0036] For example, an existing real-time customer A has a corresponding real-time customer portrait C, and the time threshold is set to 3 years. It can be obtained that the customer level of the real-time customer portrait C changes from level 1 to level 4, that is, the customer level of the real-time customer portrait C has a large degree of reduction. Therefore, when the customer relationship management of the real-time customer A is performed, the attention to the real-time customer A needs to be improved, so as to prevent the real-time customer A from also having a large degree of customer level reduction, so that the user can avoid risks in advance and improve the safety of the user.

[0037] As shown in the Figure 2 The application also provides a customer relationship management system based on artificial intelligence, which comprises a collection component 100 and a management component 200. The customer information of a plurality of customers is collected by the collection component 100. The customer relationship management method based on artificial intelligence is executed by the management component 200. The customer portrait is constructed based on the customer information by the management component 200. The customer portrait is compared with the customer standard portrait, so as to predict and adjust the customer relationship of the customer.

[0038] It should be noted that the "target customer" in the application not only refers to an individual customer, but also includes a company, a merchant and the like. After the collection of the customer information of the target customer is completed, the customer information is used as the training data of the management component 200. The customer portrait of the target customer is created based on the customer information by the management component 200, so as to manage the customer relationship of the target customer based on the customer portrait.

[0039] In an optional embodiment, the collection component 100 comprises a collection module 101 and a security module 102. The collection module 101 is used to collect the customer information of the target customer. The security module 102 encrypts the customer information, so as to prevent the leakage of the customer information.

[0040] It should be noted that, since the application is used for customer relationship management, the security of the customer data needs to be ensured when the data is collected. The collected customer information is encrypted by the encryption module 102, so as to ensure that the customer information is only used for the training of the management model, avoid the leakage of the customer information, and improve the protection of the customer information.

[0041] The embodiments of the application are described in detail above with reference to the drawings, but the application is not limited thereto. Various changes can be made within the knowledge of those skilled in the art without departing from the spirit of the application.

Claims

1. A customer relationship management method based on artificial intelligence, characterized in that: include: Collect customer information of multiple target customers and build a customer profile for each target customer based on the customer information; Create a management model; Input customer information and the corresponding customer profiles of each customer information into the management model. The management model then filters out standard customer profiles based on the customer profiles, and sets the user level of the target user based on the standard customer profiles to obtain a trained management model. Collect real-time information of real-time customers, input the real-time information into the trained management model, obtain the real-time standard profile of the real-time customers, and predict and adjust the customer relationship of the real-time customers based on the real-time standard profile.

2. The customer relationship management method based on artificial intelligence according to claim 1, characterized in that: Collect customer information of multiple target customers and build a customer profile for each target customer based on the customer information, including: Create a customer database; Collecting customer information of multiple target customers respectively, and inputting all collected customer information into a customer database; the customer information includes basic customer data and customer behavior data, the basic customer data includes customer address and customer contact information, etc., and the customer behavior data includes transaction record data and customer communication data; Combine the customer information of each target customer to build a corresponding customer portrait.

3. The customer relationship management method based on artificial intelligence according to claim 2, characterized in that: Build a corresponding customer profile based on the customer information of each target customer, including: Randomly select customer information of a target customer from the customer database; Calculate the customer value of the target customer using Formula 1 based on the customer information of the target customer; Formula 1; where is the customer value of the target customer, is the target customer’s i-th customer sub-information, is the weight corresponding to the i-th customer sub-information, and N is the total number of customer sub-information contained in the customer information; Return the customer information of a randomly selected target customer from the customer database until the customer information of all target customers in the customer database is selected, and obtain the customer value of each target customer; Set customer profiles and the customer value range corresponding to each customer profile; Assign each customer to the corresponding customer profile based on the customer value of the target customer.

4. The customer relationship management method based on artificial intelligence according to claim 3, characterized in that: Input customer information and the corresponding customer profiles of each customer information into the management model. The management model then filters out standard customer profiles based on the customer profiles and sets the user level of the target user based on the standard customer profiles. The trained management model includes: Constructing a coupling relationship between customer information and customer profile, and using customer information, customer profile, and the coupling relationship between customer information and customer profile as a training sample to obtain multiple training samples; Divide multiple training samples into training sets and test sets according to random proportions; The training samples in the training set are sequentially input into the management model. The management model then performs cluster analysis on multiple customer information under the customer profile based on the customer profile to obtain the customer standard profile corresponding to each customer profile. The customer relationship of the target customer is then set based on the customer standard profile. Input the test set into the trained management model to verify whether the trained management model is trained.

5. The customer relationship management method based on artificial intelligence according to claim 4, characterized in that: The training samples in the training set are sequentially input into the management model. The management model then performs cluster analysis on multiple customer information under the customer profile based on the customer profile to obtain the customer standard profile corresponding to each customer profile. The customer relationship of the target customer is then set based on the customer standard profile, including: For each customer profile, a training sample under the customer profile is randomly selected from the training set, and the selected training sample is recorded as the initial standard profile; Assign other training samples to the initial customer profile of the same customer profile, thereby forming multiple profile clusters; Set the iteration threshold; For each cluster, calculate the distance between each training sample in the cluster and the standard image, and record the point corresponding to the average value of the distance as the new standard image; Determine whether the number of iterations is greater than or equal to the iteration threshold; When the number of iterations is greater than or equal to the iteration threshold, the iteration is stopped and the last obtained standard portrait is recorded as the customer standard portrait, thereby obtaining the customer standard portrait corresponding to each customer.

6. The customer relationship management method based on artificial intelligence according to claim 5, characterized in that: The training samples in the training set are sequentially input into the management model. The management model then performs cluster analysis on multiple customer information under the customer profile based on the customer profile to obtain a standard customer profile corresponding to each customer profile. The customer relationship of the target customer is then set based on the standard customer profile. This also includes: Set up multiple customer levels and the corresponding customer standard profiles for each customer level; specifically, each customer level can correspond to multiple customer standard profiles; Assign corresponding customer levels based on the standard customer portrait corresponding to each target customer.

7. The customer relationship management method based on artificial intelligence according to claim 6, characterized in that: Collect real-time customer information and input it into the trained management model to obtain the real-time standard profile of the real-time customer. Combined with the real-time standard profile, predict and adjust the customer relationship of the real-time customer, including: Collect real-time information of real-time customers; Create real-time profiles of real-time customers based on real-time information; Input the real-time information and the real-time profile into the trained management model to obtain a standard customer profile corresponding to the real-time customer output by the trained management model; record the standard customer profile corresponding to the real-time customer as the real-time standard profile; Combine real-time standard profiles to predict and adjust customer relationships with real-time customers.

8. The customer relationship management method based on artificial intelligence according to claim 7, characterized in that: Combine real-time standard profiles to predict and adjust real-time customer relationships, including: Set time thresholds; Obtain the future annotated portrait corresponding to the time threshold of the real-time standard portrait; Determine whether the future standard profile and the real-time profile are the same customer profile; If the real-time standard profile and the real-time profile are not the same customer profile, the future standard profile will be assigned to the real-time customer; Set up corresponding customer relationships based on the future standard portraits given to real-time customers.

9. A customer relationship management system based on artificial intelligence, characterized in that: include: Acquisition components; Collecting customer information of multiple customers through the collection component; Management components; The artificial intelligence-based customer relationship management method described in any one of claims 1 to 8 is executed by the management component, a customer profile is constructed based on customer information by the management component, and the customer profile is compared with the customer standard profile, thereby adjusting the customer's customer relationship.

10. A customer relationship management storage device based on artificial intelligence, characterized in that: The collection component includes a collection module and a confidentiality module. The collection module is used to collect customer information of target customers, and the confidentiality module encrypts the customer information to prevent leakage of the customer information.

Images